WIP: break everything, new docs #173
Conversation
src/SolverInterface/sets.jl
Outdated
| dimension(s::Interval) = 1 | ||
|
|
||
| """ | ||
| addconstraint!(m::AbstractMathProgModel, b, a_varidx, a_coef, Q_vari, Q_varj, Q_val, S::AbstractSet) |
src/SolverInterface/sets.jl
Outdated
| ``` | ||
| where ``a`` is a sparse vector specified in tuple form by | ||
| `a_varidx`, and `a_val`; ``b`` is a scalar; | ||
| the symmetric matrix ``Q_i`` is defined by the triplets in `Q_vari`, `Q_varj`, |
src/SolverInterface/attributes.jl
Outdated
| """ | ||
| getattribute!(output, m::AbstractMathProgModel, attr::AbstractAttribute) | ||
|
|
||
| An in-place version of `getattribute!`. Return an attribute of the model `m` specified by attribute type `attr` into output vector `attr`. |
Break everything
src/SolverInterface/attributes.jl
Outdated
| """ | ||
| VariablePrimal(N) | ||
|
|
||
| The assignment to the primal variables in result `N`. If `N` is omitted, it is interpreted as 1. |
There was a problem hiding this comment.
Is the plan to implement this with a VariablePrimal() = VariablePrimal(1) constructor?
docs/src/apireference.md
Outdated
| ```@docs | ||
| AbstractAttribute | ||
| cangetattribute | ||
| getattribute |
There was a problem hiding this comment.
getattribute! is missing here
src/SolverInterface/objectives.jl
Outdated
| duplicate terms. `a_varidx`, `Q_vari`, `Q_varj` should be collections of | ||
| `VariableReference` objects. | ||
| """ | ||
| function setobjective! end |
There was a problem hiding this comment.
Worth having a method that only adds the linear part? We could have a fallback method in MPB to fill in the empty quadratic part
setobjective!(m::AbstractMathProgModel, N::Int, b, a_varidx, a_coef)
|
|
* preliminary migration of nlp tests * WIP: adding first conic example
* tests * more test * MOre testing * more tests * More tests
change free to reals, sidedim to matrix_dim, type of interval
|
I'm planning on moving this branch into a new package. Potential names: MathOptAPI (MOA) (@odow) and MathOptInterface Given that NLP is still half-baked and doesn't play well with the new design, we can leave MPB around for NLP solvers until we have a good unified API for it. |
|
You're stealing my name! Although I've been leaning more to Kōkako.jl recently... |
JuliaOpt/MathProgBase.jl#173 Includes contributions from: Site Wang <[email protected]> Joey Huchette <[email protected]> joaquimgarcia_psr <[email protected]> odow <[email protected]> ccoffrin <[email protected]> Benoît Legat <[email protected]> Chris Coey <[email protected]>
|
Development will continue here: https://github.com/JuliaOpt/MathOptInterface.jl |
Beginning of a rewrite of the docs using Documenter (#169), incorporating status changes (#164) and a new attribute concept.
We'll use this as the starting point for the documentation tutorial tomorrow.
Preview of the documentation: http://www.juliaopt.org/MathProgBase.jl/latest/